Research On Cooperative Collision Avoidance Control For Intelligent Platoon In Emergency Cut-in Scenario Under Mixed Traffic

With the increasing maturity of internet of vehicles and autonomous driving technology,the intelligent platoon coupled by intelligent connected vehicles(ICVs)has gradually become an important application of intelligent traffic system(ITS)with its advantages in improving road capacity and reducing driving energy consumption.The cooperative collision avoidance system,as a key link to ensure the safe driving of the platoon,has received extensive attention.According to ICV’s development plan,a large number of ICVs will flood into the flow of manual connected vehicles(MCVs)in the future to promote the formation of a new situation of mixed traffic flow.In this stage,the platoon will interact with MCVs and generate new risks.In particular,the driver’s emergency lane-changing behavior is sudden and unpredictable,which will cause great safety hazards to the platoon behind the target lane.Therefore,on the basis of the existing cooperative collision avoidance research,further exploration of the cooperative collision avoidance strategy for intelligent platoon in the emergency cut-in scenario through early warning and coordinated control is of great significance to promote the development of ICVs platooning.The main research work is as follows:(1)The prediction method of lane-changing behavior and collision risk assessment model in emergency cut-in scenario are proposed.For the MCV emergency cut-in scenario,a lane-changing behavior recognition model based on Gaussian mixture hidden Markov and a driving behavior prediction model based on artificial neural network are constructed.Based on the conditional random field,a collision risk assessment for the platoon and its interior is established to determine the timing of strategy execution.The early prediction of emergency lane-changing behavior provides theoretical support for the platoon to respond to unexpected situation in a timely manner.(2)A strategy for active steering collision avoidance for the vehicle in the intelligent platoon in emergency cut-in scenario is proposed.Aiming at the situation that the vehicles in the platoon cannot avoid collisions by emergency braking,a neural network PID active steering controller based on genetic algorithm optimization is designed.The Combination of Neural Network and PID control(CNNPID)is adopted to eliminate the control error.Then the advantage of genetic algorithm in global search ability is applied to optimize the weight matrix interval of neural network to further improve the control performance of steering controller,which can meet the requirements of real-time and stability in the process of emergency steering collision avoidance.(3)A Hierarchical Cooperative Collision Avoidance control(H-CCA)framework for intelligent platoon in emergency cut-in scenario is proposed.From the macro perspective,the upper layer controller proposes a method for switching collision avoidance modes for the vehicles in the platoon and a method for cluster division based on risk assessment model,and then designs two-stage cooperative collision avoidance strategy to solve the collision avoidance problem for platoon and clusters respectively.The lower system performs lateral and longitudinal motion control of the vehicle based on the calculated desired acceleration and heading angle through the control mechanism.Finally,the results of 500 simulation experiments verified that the emergency lane-changing behavior prediction method could respond 0.25 s ahead of time on average,and the collision avoidance success rate of the intelligent platoon was significantly improved under the control of the H-CCA strategy.In summary,in view of the untimely response single collision avoidance mode of the existing intelligent platoon cooperative collision avoidance strategy,this thesis improves the real-time performance,efficiency and safety of the cooperative collision avoidance strategy from behavior prediction,risk assessment,cooperative collision avoidance levels,which further expands the application scenario requirements of ITS in mixed traffic environment.